Automated personalization | Community
Skip to main content
Z
Zakk
Level 1
February 13, 2026
Question

Automated personalization

  • February 13, 2026
  • 1 reply
  • 4 views

Hi Everyone,

I am consulting a travel aggregator co (hotel + flight) and if I wish to show a personalized recommendations (of ancillaries) to all customers who have passed the booking flow after selecting the flights and entering passenger info. How can I do that? So, imagine this. First page is home page which has the booking widget. You enter info here, click on CTA, move to flight search page. Here you will have to choose your flight (second page), then on the next page, you enter personal info (name, email id, phone), then comes your ancillaries page which is where I need to show all available ancillaries plus a recommended one highlighted. 

Wanted to know if Is it possible through Recommendations and / or Automated personalization ? Suggestions if it can be little elaborate can be helpful

1 reply

Vishal_Anand
Vishal_Anand
Level 5
February 16, 2026

​@Zakk  Use Recommendations to produce candidate ancillaries, then use Automated Personalization (AP) or Target’s ranking to pick/highlight the single best recommendation per user. You can also run AP standalone to rank candidates if you already supply the catalog.

Key data & trigger:

  • Visitor identity: ECID or hashed SSO ID.
  • Context params: flight_id, fare_class, route, passenger_count, booking_stage=ancillaries.
  • Catalog feed: ancillary_id, name, price, attrs (type, refundable, bundleable).
  • Events: impression, add_to_cart, purchase (sent via at.js or server-side).

Implementation options:

  • Client-side (fast): at.js call on ancillaries page to request Recommendations/AP and render with highlighted item.
  • Server-side (secure): server calls Target decisioning, renders page with highlighted recommendation.
  • Hybrid: Recommendations → AP re-rank (best balance of candidate diversity + personalization).

Audience & trigger:
Target visitors with booking_state flag (set after passenger info submit) and fire personalization on ancillaries page load.

Measurement & fallback:

  • Conversion = ancillary clicked/added/purchased.
  • A/B test: Recommendations vs AP vs business-rule (most-popular).
  • Fallback: show highest-margin/popular ancillary if model confidence low.

Privacy & QA: Mask PII, respect consent, QA full booking flows (multi-device, logged-in/out).

Quick sample params (send with decision call):
{ booking_id, visitor_id, flight_id, fare_class, passengers, booking_stage: "ancillaries" }

Terms & ConditionsAccessibility statement

Loading footer...